Multidimensional Scaling Dimensions Research Articles

Experienced Listeners’ Perception of Timbre Dissimilarity Within and Between Voice Categories

The terms "soprano" and "mezzo-soprano" are frequently used by vocal pedagogues to describe a main category of singing timbre categorization, while the terms "lyric" and "dramatic" are often used to describe sub-categories of "soprano" and "mezzo-soprano". A handful of studies have reported on the perceptual dissimilarity of main voice categories, but few, if any, have focused on within voice category perceptual distinctions such as dramatic and lyric vocal timbre. Using stimuli collected from cisgender female singers of varying voice categories and voice weights across the pitches C4, G4, and F5, this study sought (1) to visualize an experienced listener's perception of vocal timbre dissimilarity within and between voice categories using the statistical technique of multidimensional scaling (MDS), (2) to identify salient acoustic predictors of voice category and voice weight, and (3) to determine any dependencies on pitch for the perception of vocal timbre. For the pitches C4, G4, and F5, experienced listeners (N=18) rated the dissimilarity of pairs of sung vowels produced by classically trained singers classified as follows: six mezzo-sopranos (three lighter and three heavier) and six sopranos (three lighter and three heavier). The resulting dissimilarity data were analyzed using MDS. Backward linear regression was used to see if one or more of the following variables predicted MDS dimensions: spectral centroid from 0 to 5 kHz, spectral centroid from 0 to 2 kHz, spectral centroid from 2 to 5 kHz, frequency vibrato rate, and frequency vibrato extent. Listeners also completed a categorization task where they rated each individual stimulus on two dimensions: voice category and voice weight. Visual analysis of the MDS solutions appears to show that both voice category and voice weight emerged as dimensions at pitches C4 and G4. Discriminant analysis, on the other hand, statistically confirmed both these dimensions at G4, but only voice weight at C4. At pitch F5, only voice weight emerged as a dimension, both visually and statistically. Acoustic predictors of MDS dimensions were highly variable across pitches. At the pitch C4, no MDS dimension was predicted by the acoustic variables. At pitch G4, the dimension associated with voice weight was predicted by spectral centroid from 0 to 2 kHz. At pitch F5, the dimension associated with voice weight was predicted by spectral centroid from 2 to 5 kHz and frequency vibrato rate. In the categorization task, voice category and voice weight were highly correlated at the pitches C4, G4, and when all pitches were presented together, but weakly correlated at the pitch F5. While voice category and sub-category distinctions are commonly used by singing voice professionals to describe the overall timbre of voices, these distinctions may not be able to consistently predict the perceptual difference between any given pair of vocal stimuli, particularly across pitch. Nonetheless, these dimensions do emerge in some fashion when listeners are presented with paired vocal stimuli. On the other hand, when asked to rate stimuli according to the specific labels of mezzo-soprano/soprano and dramatic/lyric, experienced listeners have a very difficult time disentangling voice category from voice weight when presented with a single-note stimulus or even a 3-note stimulus consisting of the pitches C3, G4, and F5.

Psychological responses to natural patterns in architecture

In the following experiments, we examined whether perceptions of naturalness in architecture are linked to objective visual patterns, and we investigated how natural patterns influence aesthetic evaluations of architectural scenes. Experiment 1 revealed that visual patterns of architecture explained over half of the variance in scene naturalness ratings. In Experiment 2, aesthetic preference ratings were found to relate closely to natural patterns in architecture. In Experiment 3, participants completed an image arrangement task, and multidimensional scaling (MDS) analysis was performed on the data to determine the underlying dimensions that drove scene similarity judgements. Naturalness and preference ratings both correlated strongly with MDS Dimension 1. We interpreted this dimension as representing latent perceptions of naturalistic aesthetics and found that it mediated the effects of natural patterns on scene preference. Together, these results suggest that naturalistic visual patterns may play an important role in aesthetic evaluations of architectural scenes.

Inexperienced Listeners' Perception of Timbre Dissimilarity Within and Between Voice Categories

Using stimuli produced by classically trained female singers, this study sought to examine how inexperienced listeners' ratings of timbre dissimilarity relate to singer categorization, both between and within voice category. Stimuli were produced by six mezzo-sopranos and six sopranos classically trained female voice students at three pitches, C4, G4, and F5. Within each category, mezzo-soprano and soprano, three singers were classified as lighter in timbre and three were classified as heavier in timbre. At each pitch, the 12 sung stimuli were combined in all possible pairs and presented to 56 inexperienced listeners who judged each pair for the degree of timbre dissimilarity using a visual analog scale. All stimuli also were rated by expert voice professionals in two dimensions: (1) voice category and (2) vocal weight. Multidimensional scaling (MDS) dimensions agreed with teacher-determined voice category at C4 and teacher-determined vocal weight at G4 and F5. MDS dimensions also correlated with single-note ratings of voice category and vocal weight provided by expert listeners who were unfamiliar with the singing voices used in this study. With limited exceptions, no single acoustic variable predicted the MDS dimensions that resulted from the inexperienced listeners ratings of timbre dissimilarity; however, weighted linear combinations of acoustic variables were highly predictive of numerous MDS dimensions. In spite of a lack of both range and tessitura information, timbre dissimilarity ratings provided by inexperienced listeners yielded MDS dimensions that accurately predicted voice categories and vocal weight and that highly correlated to expert listeners ratings of these two parameters.

Towards understanding speaker discrimination abilities in humans and machines for text-independent short utterances of different speech styles.

Little is known about human and machine speaker discrimination ability when utterances are very short and the speaking style is variable. This study compares text-independent speaker discrimination ability of humans and machines based on utterances shorter than 2 s in two different speaking styles (read sentences and speech directed towards pets, characterized by exaggerated prosody). Recordings of 50 female speakers drawn from the UCLA Speaker Variability Database were used as stimuli. Performance of 65 human listeners was compared to i-vector-based automatic speaker verification systems using mel-frequency cepstral coefficients, voice quality features, which were inspired by a psychoacoustic model of voice perception, or their combination by score-level fusion. Humans always outperformed machines, except in the case of style-mismatched pairs from perceptually-marked speakers. Speaker representations by humans and machines were compared using multi-dimensional scaling (MDS). Canonical correlation analysis showed a weak correlation between machine and human MDS spaces. Multiple regression showed that means of voice quality features could represent the most important human MDS dimension well, but not the dimensions from machines. These results suggest that speaker representations by humans and machines are different, and machine performance might be improved by better understanding how different acoustic features relate to perceived speaker identity.

Understanding Adolescents' Categorisation of Animal Species.

Simple SummaryWhen people try to make sense of the world they often use categorisations, which are seen as a basic function of human cognition. People use specific attributes to categorise animals with young children using mostly visual cues like number of legs, whereas adults use more comprehensive attributes such as the habitat that the animal lives in. The aim of the present study was to investigate how adolescents categorise different types of animals. A card sorting exercise in combination with a survey questionnaire was implemented. Adolescents were asked to group images of a variety of common British farm, pet, and wild animals that were printed on cards. Furthermore, adolescents were asked to rate a number of animals regarding their utility, likability, and fear, which served as affective responses. Results show that adolescents primarily use an animal’s perceived utility as a means for their categorisation along with their affective feelings towards those animals. In other words, adolescents group animals into farm, pet, and wild animals with one exception, birds. Birds, regardless of their role in society (pet, farm, or wild animal), were mostly grouped together. The results are important to understand adolescents’ perception of animals, which may explain the different attitudes and behaviours towards animals.Categorisations are a means of investigating cognitive maps. The present study, for the first time, investigates adolescents’ spontaneous categorisation of 34 animal species. Furthermore, explicit evaluations of 16 selected animals in terms of their perceived utility and likeability were analysed. 105 British adolescents, 54% female, mean age 14.5 (SD = 1.6) participated in the study. Results of multidimensional scaling (MDS) techniques indicate 3-dimensional data representation regardless of gender or age. Property fittings show that affect and perceived utility of animals explain two of the MDS dimensions, and hence partly explain adolescents’ categorisation. Additionally, hierarchical cluster analyses show a differentiation between farm animals, birds, pet animals, and wild animals possibly explaining MDS dimension 3. The results suggest that utility perceptions predominantly underlie adolescents’ categorisations and become even more dominant in older adolescents, which potentially has an influence on attitudes to animals with implications for animal welfare, conservation, and education.

Auditory multistability and neurotransmitter concentrations in the human brain.

Multistability in perception is a powerful tool for investigating sensory-perceptual transformations, because it produces dissociations between sensory inputs and subjective experience. Spontaneous switching between different perceptual objects occurs during prolonged listening to a sound sequence of tone triplets or repeated words (termed auditory streaming and verbal transformations, respectively). We used these examples of auditory multistability to examine to what extent neurochemical and cognitive factors influence the observed idiosyncratic patterns of switching between perceptual objects. The concentrations of glutamate-glutamine (Glx) and γ-aminobutyric acid (GABA) in brain regions were measured by magnetic resonance spectroscopy, while personality traits and executive functions were assessed using questionnaires and response inhibition tasks. Idiosyncratic patterns of perceptual switching in the two multistable stimulus configurations were identified using a multidimensional scaling (MDS) analysis. Intriguingly, although switching patterns within each individual differed between auditory streaming and verbal transformations, similar MDS dimensions were extracted separately from the two datasets. Individual switching patterns were significantly correlated with Glx and GABA concentrations in auditory cortex and inferior frontal cortex but not with the personality traits and executive functions. Our results suggest that auditory perceptual organization depends on the balance between neural excitation and inhibition in different brain regions.This article is part of the themed issue 'Auditory and visual scene analysis'.

Overcoming the dichotomy between open and isolated populations using genomic data from a large European dataset

Human populations are often dichotomized into “isolated” and “open” categories using cultural and/or geographical barriers to gene flow as differential criteria. Although widespread, the use of these alternative categories could obscure further heterogeneity due to inter-population differences in effective size, growth rate, and timing or amount of gene flow. We compared intra and inter-population variation measures combining novel and literature data relative to 87,818 autosomal SNPs in 14 open populations and 10 geographic and/or linguistic European isolates. Patterns of intra-population diversity were found to vary considerably more among isolates, probably due to differential levels of drift and inbreeding. The relatively large effective size estimated for some population isolates challenges the generalized view that they originate from small founding groups. Principal component scores based on measures of intra-population variation of isolated and open populations were found to be distributed along a continuum, with an area of intersection between the two groups. Patterns of inter-population diversity were even closer, as we were able to detect some differences between population groups only for a few multidimensional scaling dimensions. Therefore, different lines of evidence suggest that dichotomizing human populations into open and isolated groups fails to capture the actual relations among their genomic features.

Cerebral cortical mechanisms of copying geometrical shapes: a multidimensional scaling analysis of fMRI patterns of activation

We used multidimensional scaling (MDS) to characterize the integrative neural mechanisms during viewing and subsequently copying nine geometrical shapes. Human subjects initially looked at a central fixation point ("rest" period), then looked at a geometrical shape ("visual" period) which they copied without visual feedback ("copying" period). BOLD signal was recorded from voxels in 28 cortical areas (14 from each hemisphere) using a 4 Tesla magnet. For each voxel, signal ratios of "Visual versus Rest" (VR), and "Copy versus Visual" (CV) were calculated and used to construct two sets of Euclidean distance dissimilarity matrices for the nine shapes, with separate matrices defined for each region of interest (ROI) across subjects. The relations of perceptual and motor aspects of the shapes to MDS dimensions and specific ROIs were assessed using stepwise multiple regressions. The optimal individually scaled (INDSCAL) solutions were 2-dimensional. For the VR condition, MDS dimensions were significantly associated with the presence of crossing in a shape (Dimension 1), and with perimeter, height, cycles, peak segment speed, and horizontal symmetry (Dimension 2). ROIs most prominently associated with these dimensions essentially comprised the medial frontal lobe bilaterally, the inferior frontal gyrus bilaterally, and the left intraparietal sulcus (Dimension 1), and visual areas, including the calcarine sulcus and cuneus bilaterally (Dimension 2). These results document the expected involvement of visual areas and support the hypothesis advanced on the basis of previous findings (Lewis et al. 2003a) that a motor rehearsal of the upcoming shape copying is occurring during this visual presentation period. For the CV condition, practically one motor feature (number of segments drawn) dominated both dimensions, with a secondary engagement of horizontal symmetry in Dimension 1. The right postcentral gyrus, right intraparietal sulcus, right superior parietal lobule and right inferior parietal lobule contributed mostly to Dimension 1; the superior frontal gyrus bilaterally, right middle frontal gyrus, left postcentral gyrus, left inferior parietal lobule contributed mostly to Dimension 2; and the left superior parietal lobule and left intraparietal sulcus contributed to both dimensions approximately equally. CV BOLD activation of ROIs contributing to Dimension 1 (or to both dimensions) was significantly associated with the number of shape segments drawn. Since the direction of movement differs in successively drawn shape segments, the number of segments (minus one) equals the number of changes in the direction of movement. We conclude that this fundamental spatial motor aspect of drawing geometrical shapes is the critical variable, independent of the particular shape drawn, that dominates cortical activation during copying.

Multidimensional scaling analysis of centroid- and attack/decay-normalized musical instrument sounds

Ten sustained musical instrument tones (bassoon, cello, clarinet, flute, horn, oboe, recorder, alto saxophone, trumpet, and violin) were spectrally analyzed and then equalized for duration, attack and decay time, fundamental frequency, number of harmonics, average spectral centroid, and presentation loudness. The tones were resynthesized both with time-varying harmonic amplitudes and frequencies (dynamic case) and fixed amplitudes and frequencies (static case). Tone triads were presented to ten musically experienced listeners whose tasks were to specify the most dissimilar and most similar pairs in each triad. Based on the resulting dissimilarity matrix, multidimensional scaling (MDS) was used to position the instruments in two- and three-dimensional metric spaces. Two measures of instrument amplitude spectra were found to correlate strongly with MDS dimensions. For both the static- and dynamic-case 2-D solutions, the ratio of even-to-odd rms amplitudes correlated strongly with one of the dimensions. For the dynamic case, spectral centroid variation correlated strongly with the second dimension. Also, 2-D solution instrument groupings agreed well with groupings based on coefficients of the first two components of a principle components analysis representing 90% of the instruments’ spectral variance. [This work was supported by the Hong Kong Research Grants Council’s CERG Project 613505.]

Modeling Growth Data Using Multidimensional Scaling Profile Analysis

This paper discusses how Multidimensional Scaling (MDS) can provide an exploratory technique for identifying major growth profiles, which may be indicative of growth associated with subgroups. It briefly overviews the conventional growth models and growth mixture models, examines the assumptions related to these models, and indicates some limitations associated with these models. It then proposes an exploratory growth profile analysis using the MDS model as a complement to more specification-oriented techniques. It describes the Profile Analysis via Multidimensional Scaling model (PAMS) and extends the model for longitudinal data. The MDS profile model can solve for the growth parameters such that each MDS dimension corresponds to a major growth profile. It is argued that the MDS model provides an exploratory tool for identifying growth trends and studying individual differences with respect to those growth trends. Since MDS has not traditionally been used for longitudinal studies, the MDS growth analysis can serve as the basis for studies of the kind discussed in the paper.

The use of an auditory model in predicting perceptual ratings of breathy voice quality

Summary Despite much research, the relationship between vocal acoustic signals and perceived voice quality is not well understood. The present study used an auditory model proposed by Moore et al 10 to study how changes in the acoustic spectrum may relate to changes in perceptual ratings of breathiness. Perceptual ratings of breathiness were obtained using a multidimensional scaling (MDS) design. The stimulus distances on the dominant MDS dimension were correlated with several commonly used acoustic measures for voice quality. These distances were also compared with measures obtained from the output of the auditory model. Results show that the partial loudness of the harmonic energy obtained with the aspiration noise acting as a masker was the most important predictor of perceptual ratings of breathiness. Results also demonstrate that measures obtained from the auditory spectrum were better predictors of perceptual ratings of breathiness than were commonly used acoustic spectral measures.

Similarity and features of natural textures.

In 3 experiments the type of model that is best for conceptualizing the attentive similarity of natural textures was investigated. Different groups of participants placed pictures into groups however they wished, described the resulting clusters and multidimensional scaling dimensions, identified the objects or surfaces depicted in the pictures, and ranked the pictures along several hypothesized attribute-based dimensions. Results indicate that similarity is context dependent, that natural textures seem to be organized according to family resemblances, and that a dimensional model is inappropriate. These outcomes suggest that models of preattentive segregation and attentive cognition may be incommensurable.

Similarity and features of natural textures.

In 3 experiments the type of model that is best for conceptualizing the attentive similarity of natural textures was investigated. Different groups of participants placed pictures into groups however they wished, described the resulting clusters and multidimensional scaling dimensions, identified the objects or surfaces depicted in the pictures, and ranked the pictures along several hypothesized attribute-based dimensions. Results indicate that similarity is context dependent, that natural textures seem to be organized according to family resemblances, and that a dimensional model is inappropriate. These outcomes suggest that models of preattentive segregation and attentive cognition may be incommensurable.

The dimensions of positive emotions.

This study was designed to investigate the dimensionality of positive emotions and was run in two parts. In part 1, subjects were given a set of 24 different life-situations all of which commonly give rise to positive emotions. They sorted the 24 situations into groups on the basis of whether or not they produced similar emotional responses. In part 2 of the study subjects rated each of the 24 situations on 13 emotion scales. The grouping data obtained in part 1 were submitted to multidimensional scaling (MDS) and returned a four-dimensional solution. Canonical correlations between the four MDS dimensions and the 13 emotion scales revealed that dimension 1 is best explained by 'absorption', dimension 2 by 'potency', dimension 3 by 'altruistic' and dimension 4 by 'spiritual'. These correlations were then married to an interpretation of the situations falling high and low on each of the four dimensions, with the following results. Dimension 1 distinguishes internal or private situations from social situations, dimension 2, achievement from leisure situations, dimension 3, social demands from self-indulgence, and dimension 4, serious from trivial situations. The implications of these findings for the study of emotion are discussed.

A multidimensional scaling study of normal voice quality in females

Little is currently known about the perceptual dimensions listeners use in discriminating female voices and how these dimensions relate to perceptual rating‐scale judgements. The purpose of this study was to examine these issues for normal female voices. In addition, the judgments of 20 naive listeners and 20 experienced listeners (speech pathologists) were compared. Listeners provided dissimilarity judgments of sentences produced by 20 females (24–25 years). The listeners also rated each voice on 17 perceptual rating scales (e.g., high pitch‐low pitch, clear‐hoarse). Results of the multidimensional scaling (MDS) analysis produced a five‐dimensional solution (RSQ = 0.509). A stepwise linear regression related the perceptual attributes of pitch, age, loudness, “liveliness,” and pitch variability/quality to the five MDS dimensions. When listener groups were compared, the experienced listeners relied on pitch somewhat more than the naive listeners, while the naive listeners utilized the pitch variability/quality dimension to a greater degree. This study demonstrates that listeners' perceptual ratings can be significantly correlated with perceptual dimensions derived from MDS analysis. [Research supported by BRSG Grant S07 RR07031, DRR, NIH.]

Individual Differences in Perceptions of Occupations and Occupational Reinforcers

Individual differences in subjective occupational perceptions were investigated. Forty‐three students and 11 vocational counsellor/researchers judged the similarity of all possible pairs of 18 occupations, three from each of Holland's categories. Subjects rated each occupation on 11 occupational reinforcers. The unity of all but Holland's Conventional category was supported by the configuration of occupations in a four‐dimensional representation obtained from a multidimensional scaling (MDS) of the similarity judgements. Holland's hexagonal arrangement of the categories was not supported. Regressing occupational reinforcer ratings on to MDS dimensions indicated that the reinforcers were generally a significant component of occupational perceptions. The Independence reinforcer, however, was a significant component only for the counsellor/researchers and Compensation was more salient to the students than to the counsellor/researchers.

Foundations of multi‐dimensional metric scaling in Cayley‐Klein geometries

Two‐dimensional metric scaling is based on the relations of incidence and orthogonality and on the motion of reflection in a straight line. Together with a set of optional postulates concerning different types of orthogonality these primitives are sufficient for operationalizing the Cayley‐Klein geometries for use in the psychological laboratory. They include the three Riemannian geometries (elliptic, Euclidean and hyperbolic) as special cases. The representation refers to a vector space over an unspecified field. The advantage of this lies in the identifiability of the geometry present, before technical axioms have been invoked in order to secure a field of real numbers.The approach appears to be commendable because it yields new solutions to classical problems from monocular and binocular space perception as well as from colour vision. Extensions to derived measurement outside of psychophysics and to three and more dimensions of multidimensional scaling are reported.